Reopening openness to experience: A network analysis of four openness to experience inventories By: Alexander P. Christensen, Katherine N. Cotter, and Paul J. Silvia Alexander P. Christensen, Katherine N. Cotter, & Paul J. Silvia (2018). Reopening openness to experience: A network analysis of four openness to experience inventories. Journal of Personality Assessment. This is an Accepted Manuscript of an article published by Taylor & Francis in Journal of Personality Assessment on 10 May 2018, available online: http://www.tandfonline.com/10.1080/00223891.2018.1467428 ***© 2018 Taylor & Francis. Reprinted with permission. No further reproduction is authorized without written permission from Taylor & Francis. This version of the document is not the version of record. Figures and/or pictures may be missing from this format of the document. *** Abstract: Openness to Experience is a complex trait, the taxonomic structure of which has been widely debated. Previous research has provided greater clarity of its lower order structure by synthesizing facets across several scales related to Openness to Experience. In this study, we take a finer grained approach by investigating the item-level relations of four Openness to Experience inventories (Big Five Aspects Scale, HEXACO–100, NEO PI–3, and Woo et al.'s Openness to Experience Inventory), using a network science approach, which allowed items to form an emergent taxonomy of facets and aspects. Our results (N = 802) identified 10 distinct facets (variety-seeking, aesthetic appreciation, intellectual curiosity, diversity, openness to emotions, fantasy, imaginative, self-assessed intelligence, intellectual interests, and nontraditionalism) that largely replicate previous findings as well as three higher order aspects: two that are commonly found in the literature (intellect and experiencing; i.e., openness), and one novel aspect (open- mindedness). In addition, we demonstrate that each Openness to Experience inventory offers a unique conceptualization of the trait, and that some inventories provide broader coverage of the network space than others. Our findings establish a broader consensus of Openness to Experience at the aspect and facet level, which has important implications for researchers and the Openness to Experience inventories they use. Keywords: openness to experience | personality inventories | network science Article: Openness to Experience is a broad and complex trait that has gone by many names over the years, such as Openness to Experience, Intellect, Culture, Imagination, and Creativity (Fiske, 1949; Goldberg, 1981; Johnson, 1994; Norman, 1963; Saucier, 1992). Given the trait's breadth and complexity, researchers have identified two aspects of Openness to Experience: Openness to Experience (for clarity, hereafter referred to as Experiencing, following Connelly, Ones, Davies, & Birkland, 2014) from the questionnaire tradition (Costa & McCrae, 1992), and Intellect from the lexical tradition (Goldberg, 1981). The experiencing aspect is characterized by an appreciation for aesthetics, openness to emotions and sensations, absorption in fantasy, and engagement with perceptual and sensory information (DeYoung, Grazioplene, & Peterson, 2012). The intellect aspect is characterized by intellectualism, enjoyment of philosophy, curiosity, and engagement with abstract and semantic information (DeYoung et al., 2012). Beneath the Experiencing and Intellect aspects, however, are many lower order facets of Openness to Experience. The measurement of these facets has been inconsistent, with some facets being measured in some inventories but not in others. As a result, this has led to variation in the coverage and conceptualization of the Openness to Experience construct. Despite research examining the content and number of facets (Connelly, Ones, Davies, et al., 2014; Woo et al., 2014), there has yet to be an empirical investigation into how the items of different inventories converge (or diverge) on the coverage and content of lower order facets. Therefore, in this research, we sought to clarify how four commonly used Openness to Experience inventories conceptualize the construct. In addition, we wanted to clarify the number and content of the lower order facets across these inventories. To do so, we applied a computational network science approach to construct a network using the items from these four inventories. From this network, we used a community detection algorithm to identify communities (i.e., facets) in the network. These network-identified facets were then used to examine the conceptual coverage of each inventory—whether items of the inventory were represented in many or a few of the network-identified facets. Openness to experience taxonomy Past debates about how the global Openness to Experience trait should be defined has subsided—traditional factor analysis approaches have identified both experiencing and intellect as aspects of the global trait (DeYoung, Quilty, & Peterson, 2007; Woo et al., 2014). Experiencing and intellect are separate but related aspects of Openness to Experience, with differential relations to affective, behavioral, and cognitive outcomes (Barford & Smillie, 2016; DeYoung et al., 2012; DeYoung et al., 2014). For example, experiencing is positively related to creative achievement in the arts (Kaufman et al., 2016), implicit learning (Kaufman et al., 2010), and feeling mixed emotions (Barford & Smillie, 2016), whereas intellect is positively related to creative achievement in the sciences (Kaufman et al., 2016), working memory (DeYoung, Shamosh, Green, Braver, & Gray, 2009), and fluid intelligence (DeYoung et al., 2012). As a result, these aspects have been generally agreed on. Beneath the experiencing and intellect aspects, however, the lower order facet structure of Openness to Experience becomes less clear—depending on which inventory is used, the number of facets included can range from four to nine (Connelly, Ones, & Chernyshenko, 2014). There appears to be some level of agreement on the importance of some facets (e.g., aestheticism, intellectualism) because they are featured in many inventories. Many facets, however, are unique to only one or two inventories (e.g., Feelings, Actions, Curiosity). Additionally, some inventories seem to provide good coverage of the facets in one aspect but have limited coverage of facets in the other. For instance, in the Woo et al. (2014) factor analysis of seven Openness to Experience inventories, facets of the Revised NEO Personality Inventory (NEO PI–R) loaded onto the experiencing aspect well but had relatively low loadings for the intellect aspect. Thus, although several Openness to Experience inventories exist, it appears that they vary substantially in their coverage of the trait's conceptual space. The special section in the Journal of Personality Assessment sought to reach a broader consensus of this lower order taxonomy (Connelly, Ones, & Chernyshenko, 2014). Connelly, Ones, Davies, et al. (2014) undertook the most comprehensive theoretical evaluation of Openness to Experience's lower order facets to date by theoretically sorting and meta-analyzing 85 Openness to Experience–related scales. They identified 11 facets that were theoretically and empirically related to Openness to Experience: Aestheticism, Autonomy, Fantasy, Innovation, Introspection, Nontraditional, Openness to Emotions, Openness to Sensations, Thrill-seeking, Tolerance, and Variety-seeking. Only four of these facets, however, were considered pure (i.e., not related to any other personality trait): aestheticism, openness to sensations, nontraditional, and introspection. Based on Connelly et al.'s sort, these pure facets were placed within the experiencing (aestheticism and openness to sensations) and intellect (nontraditional and introspection) aspects. The other seven facets aligned with Openness to Experience and other personality traits, so they were labeled as trait compounds. For example, fantasy was positively associated with Openness to Experience and negatively with Conscientiousness, whereas innovation, openness to emotions, thrill-seeking, and variety-seeking were positively associated with Openness to Experience and Extraversion. In sum, their extensive analysis of Openness to Experience–related scales provides a general framework for defining which facets are central to the construct. In the same special section, Woo et al. (2014) empirically evaluated Openness to Experience's lower order structure by factor analyzing a multitude of openness-related scales and assembling a comprehensive inventory. Prior to inventory construction, Woo et al. synthesized several taxonomic approaches to inventory development—questionnaire, lexical, and subject matter experts—to systematically organize their measurement model of Openness to Experience. They used an exploratory factor analysis (EFA) on the existing facets of 36 openness-related scales to uncover two aspects and six facets (Table 1). Subject matter experts reviewed the content (i.e., original facets and their items) of the facets identified in the factor analysis to generate conceptual definitions for these six facets. From these conceptual definitions, a shortened inventory was developed that produced a 54-item Openness to Experience inventory, which was then examined in cross-cultural samples (Woo et al., 2014). In drawing from other inventories, Woo et al. began to establish a more comprehensive lower order facet structure of Openness to Experience. Connelly, Ones, Davies, et